Interobserver Agreement in Case History Evaluation in Carpal Tunnel Syndrome

Abstract

Objectives: 

The case history is the cornerstone for the diagnosis of carpal tunnel syndrome (CTS). The value of neurological examination in CTS seems limited. In this study, we investigated the interobserver agreement in case history of CTS and the potential additional value of neurological examination.

Methods: 

Case history was taken and neurological examination was done in all patients referred for a possible CTS during a 6-month period. This was done independently by 2 senior neurologists.

Results: 

We studied 119 patients. The interobserver agreement for the conclusion of CTS after history was high (k = 0.79). Neurological examination rarely changed the diagnosis based on the case history.

Conclusions: 

This study shows a high interobserver agreement in the diagnosis of CTS based on patient history alone. Neurological examination rarely changed this diagnosis, suggesting a prominent role for the case history and a limited role for neurological examination. This may have practical implications.

INTRODUCTION

Carpal tunnel syndrome (CTS) is the most common peripheral nerve disorder, accounting for 90% of all neuropathies, caused by compression of the median nerve at the wrist. CTS patients share a characteristic case history.^1–4 Typically, the patient wakes up during the night because of tingling and sometimes pain in the palmar side of the hand and the fingers supplied by the median nerve. During daytime, most patients complain of tingling in the same area, usually provoked by performing activities with flexed wrists, for instance, riding a bicycle or driving a car. Numbness of 1 or more fingers and weakness or clumsiness of the hand may also be present.^1,4 The complaints may diminish or disappear by shaking the hand (flick sign).

Although there is no gold standard, CTS is generally diagnosed by combining findings in case history, neurological investigation, and nerve conduction studies (NCS). However, the value of neurological examination in CTS is questionable.^3,5,6 Sensory disturbances are often found but are subjective and not always confined to the area of the median nerve. Muscle weakness and atrophy of the hand muscles supplied by the median nerve are rarely present.⁵ Furthermore, maneuvers to provoke the symptoms by flexing the wrist (Phalen⁷), or tapping the median nerve at the wrist (Tinell⁸), lack specificity.^{3, 9}

NCS are often abnormal, the sensitivity is estimated to be about 85% and the specificity is more than 95%.¹⁰ Taylor-Gjevre et al¹¹ found much lower sensitivity and specificity in patients with a high pretest probability of CTS. This implies that a number of patients with classical symptoms may have normal NCS and that subjects without any complaints may have abnormal NCS.

Because of the above-mentioned disadvantages, patient history is generally considered to be the most important factor in the diagnosis of CTS.¹² If history taking according to a standardized method is reliable with a high interobserver agreement and if additional neurological investigation proves to be redundant, then evaluation of CTS patients can be accomplished much more efficiently.

In this study, we developed a standardized method of history taking based upon recent reviews about CTS.^1,12,13 We investigated the interobserver agreement in the diagnosis of CTS using this standardized history taking. Furthermore, the potential additional value of a neurological examination and NCS was evaluated.

MATERIALS AND METHODS

Patient Population

During a period of 6 months, 159 consecutive patients were evaluated for participation in this study. They were referred by their general practitioners to the neurological outpatient clinic with a possible diagnosis of CTS. Patients were excluded if they suffered from CTS before, if complaints had disappeared before the visit to the clinic, or if they were unable to speak Dutch fluently. In accordance with Dutch law and after consultation with the ethics expert of the Albert Schweitzer Hospital, we informed all referred patients about this study. All included subjects gave informed consent.

Standardized History Taking and Examination

Each patient was assessed by 2 of 3 observers (D.W., H.K., or R.P.K.). We developed a standardized questionnaire based on recent literature, in which the most common complaints of CTS are described (Table 1).^1,12,13 We asked for pain and paresthesia in the hand and fingers (especially in the distribution of the median nerve) and whether there was nocturnal paresthesia.

TABLE 1: The Standardized History and Physical Examination

Based on the results of the history, the 2 observers independently determined the likelihood of the presence of CTS. Because there is no gold standard, it was left to the clinical judgment of the examiners to come to a conclusion. We used 4 categories (no CTS, uncertain CTS, probable CTS, and certain CTS). Both observers then performed a neurological examination (Table 1). Muscle strength was graded according to the medical research council scale, and sensibility was tested by light touch and pinprick on the fingers and hand. The interobserver agreement was calculated. We also studied how often the initial diagnosis, based on the history, was changed after the neurological examination. For this, we used the decisions of the first observer.

Nerve Conduction Study

In all patients, NCS was performed. We measured the distal sensory latency of the sensory median and ulnar nerve, stimulating at the wrist and recording antidrome from the fourth digit with an equal distance between the stimulation site of the median and ulnar nerve and the recording place. Furthermore, we measured the distal motor latency of the median and ulnar nerve, stimulating the median nerve at the wrist and recording from the second lumbrical muscle and stimulating the ulnar nerve at the wrist and recording from the second palmar interosseous muscle, with an equal distance between the stimulating places and the recording place. A difference between the distal sensory latency and the distal motor latency greater than 0.4 milliseconds of the median nerve relative to the ulnar nerve was considered to be abnormal.^10,14 If only the distal sensory latency or the distal motor latency were abnormal, a third test was performed. For this test, the radial and median nerves were stimulated at the wrist with antidrome recording of the sensory response at the first digit, again with an equal distance between the stimulating places and recording place. Finally, the median nerve was considered abnormal and supporting the diagnosis of CTS if 2 out of the 3 performed tests were abnormal. The NCS was used to draw the conclusion “yes” or “no” CTS. In all patients, ulnar motor and sensory nerve studies were normal. The results of the nerve conduction study were compared with the prediction of the observers based on the case histories.

Statistics

Statistical analysis was performed using the statistical software program SPSS version 16. The k was calculated for the variation in interpretation based on the patient's medical history between 2 researchers. To do this we reduced the number of categories in which the patients were classified from 4 to 3. The categories uncertain CTS and probable CTS were combined to form the category CTS diagnosis uncertain.

RESULTS

Patients

From February to July 2009, 159 patients were evaluated who had been referred by general practitioners suspected of CTS. Of these, 119 patients (73%) were included in the study. Forty patients (26%) were excluded (Table 2). The main reason for exclusion was relapsing CTS (57.5% of the excluded patients). The male-to-female ratio was 1:4 and mean age was 51.5 years (range, 21–84 years). On average, the complaints persisted for almost 2 years, ranging from 2 weeks to 20 years. The most common symptoms in the population studied were tingling during the night (84%), tingling during daytime (86%), tingling related to exercise (82%), and pain (71%).

TABLE 2: Excluded Patients

Interobserver Agreement

Based on the history, patients were classified into 4 categories, each researcher blinded for the result of the other (no CTS, uncertain CTS, probable CTS, and certain CTS). Next, these 4 categories were reduced to 3: no CTS, certain CTS, and uncertain CTS (this category combines the categories uncertain CTS and probable CTS). Comparing these 3 diagnostic categories, the calculated κ is 0.79, indicating a good agreement. Table 3 shows the agreement between the 2 observers.

TABLE 3: Interobserver Agreement in CTS History Taken by 2 Observers.

The Value of Physical Examination in Decision Making

We also determined the value of a physical examination in addition to the history, using the results of the first observer. In only 2 of 119 cases the initial diagnosis based on the history was changed from uncertain diagnosis to “certain” CTS.

The Value of NCS in Diagnosing CTS

Finally, we examined the additional value of NCS in the various history-based categories (no CTS, uncertain CTS, and certain CTS). The results are shown in Figure 1. Within the category certain CTS, this conclusion was confirmed by NCS in 94% of the patients, 5 patients (6%) had normal NCS results.

FIGURE 1: Flow chart of the distribution of the case history, physical examination, and NCS in patients with suspected CTS.*

Within the category no CTS, an abnormal median nerve conduction was found in 2 of 9 patients (22%). Within the category uncertain CTS, 57% of the patients had an abnormal NCS of the median nerve. Because there is no independent gold standard for diagnosing CTS, it is in fact impossible to determine which patients in this category suffer from CTS.

DISCUSSION

The case history is often reported to be the cornerstone in the diagnosis of CTS.¹³ The observation that the majority of the patients referred by general practitioners with probable CTS indeed had a CTS may support this. If so, it is important that a conclusion based on the history should be reliable. We investigated the interobserver agreement based on case histories of patients with suspected CTS. We found a high level of agreement (k) between 2 investigators (0.79) if a standardized questionnaire was used.

Our study also showed that neurological examination rarely changed the history-based diagnosis. Physical examination only had additional value if the history-based diagnosis was uncertain CTS.

Furthermore, our study shows a high correlation between the findings of standardized history and NCS in patients with clinically certain CTS. This is according to Jablecki et al¹⁰ but in contrast to Taylor-Gjevre et al.¹¹ The additional value of NCS therefore remains unclear. A normal median nerve conduction in these clinically certain CTS patients might support a decision for an initially more conservative therapeutic approach. In the group of patients where the history-based diagnosis is uncertain, NCS can be useful to support or reject the diagnosis CTS. Especially in these groups (2/119 and 12/119, respectively), NCS influenced patient management. For the groups with a clinically certain negative diagnosis, results of NCS are unlikely to influence the further approach of a clinician. In most cases, no additional NCS would be performed in these patients. Especially in the group of patients with a history-based conclusion of certain CTS but a normal NCS, more scientific research is recommended, but these studies are difficult because no true gold standard for CTS exists. Perhaps ultrasound examination will be helpful in these cases.¹⁵ A clinically important question is whether the findings of median NCS in these cases have additional predictive value concerning the outcome of treatment approaches.

In conclusion, our present study shows a high interobserver agreement when taking a case history of CTS using a standard questionnaire. Additional physical examination is of limited value in the diagnosis of CTS in patients with a suspected CTS and of no value in patients with a certain diagnosis. NCS are especially useful in patients where the history-based diagnosis of CTS is uncertain. The clinical meaning of conflicting results between history and NCS still remains unclear. Our observation supports the idea that the history in patients with CTS is the cornerstone of the diagnostic procedure. This has great practical implications for routine daily practice. A brief interview by phone by a trained nurse may be enough to schedule the patient for an NCS and a short visit to the neurologist afterward. It even may be justified to skip the physical examination in typical patients.